Professor of Cell Biology

Harvard Medical School
Armenise 625B
200 Longwood Avenue
Boston, MA 02215
Tel: (617) 5829717
Email: and


Lab size:Between 5 and 10

The overall goal of our work is to solve biological problems using quantitative methods from the physmathstatdata sciences. We use computational methods from statistical physics (e.g. maximum entropy, inference of probability distributions), data sciences andstatistics (e.g. machine learning), as well as computer science and mathematics (e.g. network analysis, optimization methods). We apply these methods to build predictive network models of molecular and cellcell interactions, to support cancer precision medicine, and to aim for discoveries in structural and evolutionary biology.

Predictive network models : In a wet and dry lab, and in collaborations, we are developing systems biology methods that combine systematic perturbation experiments with rich observational readout, for the de novo derivation of predictive and quantitative models. We aim to model cellular singling events, e.g., between proteins, nucleic acids and metabolites, and cell-cell interactions in cell culture or cancer tissues. To solve the computational complexity of systematic model discovery we adapt methods from statistical physics and data sciences. The translational impact of this technology is in the nomination of combination therapeutics in cancer, such as melanoma, prostate cancer and pancreatic cancer. The cell biological impact depends on the availability of rich perturbation data with rich readout, with dual Crispr screens followed by mass spectrometry an attractive next technological advance and using other datasets from the HMS community may be an exciting opportunity.

Cancer precision medicine: Bioinformatics took on a crucial analytical and infrastructure role in the national and international cancer genome atlas projects. Collaborating closely with clinical
researchers, we are applying bioinformatics analysis to cancer genomics, for individual cancer types and across the board. For example, pancancer oncogenic signature analysis of about 10K
tumor samples provides an estimate as to which groups of cancer patients can be nominated for particular trials based on shared occurrence of genomic alterations. Collaboration with clinicians for the design of ‘basket' and ‘match' trials would be attractive.

Structural and evolutionary biology: In collaboration with the Marks group ins systems biology at HMS, we are interested in strengthening the application of maximum entropy methods, with improved inference algorithms, to problems of evolutionary biology, structural biology and cell biology. The derivation of interactions that explain the observed correlations in data is key and there is significant potential for inferring interactions in diverse biological systems beyond proteins and RNA. We would like to generate evolutionarily constrained sequences in the laboratory and further develop a quantitative theory relating biopolymer sequences to phenotypic consequences. General goals and style of science: In general terms, I am in favor of choosing problems that have a chance of impacting the real world, such as human health, global future, synthetic biology, and engineering. Working with one's group and with collaborators, I tend to have in mind three goals to guide decisions: focus on the common goal of scientific progress, build and practice supportive and constructive human relations, and help with everyone’s career goals all three, and in that order of priority. Within the community, I also would like to promote creative ways of communicating scientific results and to help reform the counterproductive journal publication system.

Collaborations : There are a number of collaborative opportunities with SysBio and CellBio faculty and I look forward to actively exploring lean, interesting, effective and consequential collaborative projects. As a declared and enthusiastically accepted goal when I started at DFCI and HMS, I plan to promote translational work bridging basic quantitative science to problems of human health, especially cancer and hope to help build more bridges between the two institutions DFCI and HMS.